AI in Simple Terms: Predicting What Comes Next
In this Short, I strip the explanation of a language model back to prediction.
Given the context in front of it, a model produces a continuation using patterns learned during training. “Guessing” is a useful everyday entry point, although the machinery behind that prediction is much more involved.
The distinction I want to keep is between a plausible continuation and a verified fact. A sentence can fit the pattern beautifully and still say something untrue.
Try noticing that difference with an ordinary question. Read the answer, then separate its wording from the evidence behind its claims. Which parts can you check?
The Short is a starting point for understanding the behavior, not a complete account of model training. The longer dot-and-line explanation develops the picture further.